L-4: Difference between revisions

From Disordered Systems Wiki
Jump to navigation Jump to search
Line 10: Line 10:


==Let's do replica!==
==Let's do replica!==
To make progress in disordered systems we have to go through the moments of the  partition function. We recall that  
To make progress in disordered systems we have to go through the moments of the  partition function. For simplicity we consider polymers starting in <math>x_0=0</math> and ending in  <math>x_t</math>. We recall that  
* <math>V(x,\tau)</math> is a Gaussian field with
* <math>V(x,\tau)</math> is a Gaussian field with
<center> <math>
<center> <math>
Line 21: Line 21:
The first moment of the partition function is
The first moment of the partition function is
<center> <math>
<center> <math>
\overline{Z_t[x_1] } =\int {\cal D} x_1 \exp\left[- \frac{1}{T} \int_0^t d \tau \frac{1}2(\partial_\tau x)^2\right]  \overline{\exp\left[- \frac{1}{T} \int d \tau V(x,\tau ) \right]}
\overline{Z_t[x_t,t] } =\int_{x(0)=0}^{x(t)=x_t} {\cal D} x_1 \exp\left[- \frac{1}{T} \int_0^t d \tau \frac{1}2(\partial_\tau x)^2\right]  \overline{\exp\left[- \frac{1}{T} \int d \tau V(x,\tau ) \right]}
</math></center>
</math></center>
Note that the term <math> T^2 \overline{W^2} = \int d \tau_1 d\tau_2 \overline{V(x,\tau_1)V(x,\tau_2)}= D t \delta(0)</math> has a short distance divergence due to the delta-function.  Hence we can write:
Note that the term <math> T^2 \overline{W^2} = \int d \tau_1 d\tau_2 \overline{V(x,\tau_1)V(x,\tau_2)}= D t \delta(0)</math> has a short distance divergence due to the delta-function.  Hence we can write:
<center> <math>
<center> <math>
\overline{Z_t[x_1] } = \frac{1}{\sqrt{2 \pi t T}}\exp\left[ -\frac{x^2}{2t T} \right]  \exp\left[ \frac{D  t \delta(0)}{2T^2}  \right]
\overline{Z_t[x_1] } = \frac{1}{\sqrt{2 \pi t T}}\exp\left[ -\frac{x_t^2}{2t T} \right]  \exp\left[ \frac{D  t \delta(0)}{2T^2}  \right]
</math></center>
</math></center>



Revision as of 15:31, 5 January 2024

Goal 1: final lecture on KPZ and directed polymers at finite dimension. We will show that for d>2 a "glass transition" takes place.

Goal 2: We will mention some ideas related to glass transition in true glasses.


Part 1: KPZ in finite dimension

  • In d=1 we found θ=1/3 and a glassy regime present at all temperatures. Moreover, the stationary solution tell us that Emin[x] is a Brownian motion in x. However this solution does not identify the actual distribution of Emin for a given x. In particular we have no idea from where Tracy Widom comes from.
  • In d>1 the exponents are not known. There is an exact solution for the Caley tree (infinite dimension) that predicts a freezing transition to an 1RSB phase (θ=0).

Let's do replica!

To make progress in disordered systems we have to go through the moments of the partition function. For simplicity we consider polymers starting in x0=0 and ending in xt. We recall that

  • V(x,τ) is a Gaussian field with
V(x,τ)=0,V(x,τ)V(x,τ)=Dδ(xx)δ(ττ)
  • From the Wick theorem, for a generic Gaussian W field we have
exp(W)=exp[W+12(W2W2)]

The first moment of the partition function is

Zt[xt,t]=x(0)=0x(t)=xt𝒟x1exp[1T0tdτ12(τx)2]exp[1TdτV(x,τ)]

Note that the term T2W2=dτ1dτ2V(x,τ1)V(x,τ2)=Dtδ(0) has a short distance divergence due to the delta-function. Hence we can write:

Zt[x1]=12πtTexp[xt22tT]exp[Dtδ(0)2T2]

Exercise L4-A: the second moment

  • Step 1:
Z[xt,t]2=exp[Dtδ(0)T2]𝒟x1𝒟x2exp[0tdτ12T[(τx1)2+(τx2)2]+1T2δ[x1(τ)x2(τ)]]

Now you can change coordinate X=(x1+x2)/2;u=x1x2 and get:

Z[xt,t]2=(Z[xt,t])2u(0)=0u(t)=0𝒟uexp[0tdτ14T(τu)2+1T2δ[x1(u(τ)]]

Part 2: Structural glasses